Significance of the Research
- We are also pursuing new approaches to weather information analysis using large language models (LLMs).
- For example, by automatically generating natural weather forecast texts from numerical weather data, we translate specialized forecast information into language that is easier for people to understand. In addition, we are exploring “human sensing” by analyzing large volumes of text data posted on social media and other platforms, in order to capture how people perceive weather conditions and what kinds of weather-related concerns attract their attention.
- In this way, by using LLMs to bridge numerical data and natural language, we aim to advance conventional weather forecasting toward a more human-centered and interactive framework, and to establish a next-generation weather information platform that supports societal understanding, decision-making, and action.
Weather Forecast Text Generation Using Multimodal LLMs
Precipitation Estimation Using Embedding Representations from Language Models
(last update: 2026/04/08)